Tesis doctoral de Chuanhe Lu
Carbon steel and compacted bentonite have been proposed as candidate materials for overpack and buffer, respectively, in a multi-barrier system of a deep geological repository for high level radioactive waste (hlw). Corrosion of the carbon steel will affect the lifetime of the overpack, the chemistry of the bentonite porewater and radionuclide migration from the canister to the host rock through the bentonite. The assessment of the performance of a radioactive waste repository requires the use of models for the long-term prediction of radionuclide migration through the near- and far-field barriers. These models must consider simultaneously the geochemical and hydrodynamic processes. Therefore, a coupled approach is needed to assess the combined effect of geochemical and hydrodynamic processes. the creation of lakes in former surface mines has been proposed in many places as an environmentally acceptable solution for the reclamation of mined lands. The water quality of a mine lake may deteriorate during lake formation due to acidification and metal pollution. The management of mine water and the evaluation of remediation actions require the use of models for understanding the natural processes involved in pollution generation and for the long-term prediction of lake water quality. reactive transport modeling provides a tool for comprehensive, quantitative and predictive treatment of chemical reactions and mass transfer within a geosystem. Reactive transport models help to understand a system as a whole and allow the quantification of the fate of contaminants through natural and engineered barriers. They have had a significant impact on the treatment of contaminant retardation in the subsurface and the description of chemical fluxes between major earth reservoirs. this dissertation deals with the development, improvement, verification and application of reactive transport models to radioactive waste disposal and acid mine drainage and lakes. The safe disposal of radioactive waste in deep geological repositories and the creation of artificial lakes in open pit mines are at the top of the current list of environmental challenges. The research group of udc led by professor javier samper is working on the application of numerical models for surface and groundwater quality in both of these problems. This dissertation has benefited from research projects of both fields and has also contributed to such projects. The geochemical models used for hlw disposal and acid mine lakes share: 1) the complexity of the chemical processes controlling the geochemistry of the system. 2) the relevance of surface chemistry processes. Bentonites are known to have a large sorption capacity with complex surface phenomena. Such surface processes occur also at the edges of fe oxi-hydroxides which form in the acid mine drainage waters. core2dv4 (samper et al., 2003) is a reactive transport code for chemically reactive non-isothermal flow of multiphase fluids in porous and fractured media. During the course of this dissertation the code has been extensively used, tested, debugged, and improved. The code has been updated to account for canister corrosion in a more realistic manner than that of standard models. The updated version of core2dv4 can simulate the progressive corrosion by assuming that the corrosion front penetrates into the canister as the corrosion progresses until all the canister has been corroded. Core2dv4 has been updated to deal with multisite surface complexation and cation exchange based on the most recent developments on thermodynamic sorption models (tsm). These updates have been verified with the experimental data obtained by bradbury and baeyens (1997) for ni2+ and poinssot et al. (1999) for cs+. Other significant improvements of core2dv4 carried out in this dissertation are: 1) including more flexibility in the choice of tsm options to allow for the possibility of considering or not the electrical terms on surface complexation (non-electrical models); 2) implementing a connection between the concentration of sorption sites and the amount of precipitated oxide minerals which contain chemically-reactive surfaces. This improvement has been motivated by the need to model the surface complexation of protons in acid mine waters in which iron oxides, hydroxides and oxy-hydroxides may precipitate as suspended matter in water; 3) improving the numerical methods to solve the nonlinear geochemical reactions by including the concentration of sorption sites s-oh within the list of unknowns in the newton-raphson iterative method. This improvement has resulted in a reduction of numerical and mass balance errors in several case studies; 4) updating and improving the calculation of the mass balance of aqueous species and water; 5) verifying the coupled transport of chemical species and stable and radioactive isotopes; 6) improving the output of sorption results; and 7) correcting some bugs in the source code. All these improvements and corrections have been verified with existing experiments and simulations. the updated version of core2dv4 has been used for reactive transport modeling of open pit mine lakes and radionuclide migration for hlw repositories within the framework of two projects funded by the european commission in the euratom program (nfpro and pamina). The first case of radioactive waste disposal has involved the modeling of the long-term geochemical evolution of a spent-fuel carbon-steel canister repository in a granite formation. Simulations have been performed with 1-d and 2-d finite element grids. They provide similar results, but the 1-d model requires half of the cpu time. The water flux through the granite rock and the excavation damaged zone around the bentonite has a significant influence on the geochemical evolution. Canister corrosion has been modeled in two stages. In the first one all chemical reactions except corrosion have been assumed at local chemical equilibrium. Model results indicate that magnetite is the main corrosion product. Fe diffusion from the canister into the bentonite leads to magnetite precipitation in the bentonite. Siderite also precipitates, but its precipitated amount is small due to the limited availability of dissolved carbonates. Corrosion causes an increase of ph in the bentonite. Proton surface complexation is the main ph buffering mechanism. Magnetite precipitation causes a reduction of porosity in the bentonite. The results of the sensitivity analysis indicate that the ph of the bentonite and the precipitation of magnetite are sensitive to the corrosion rate. The second stage of the model of the corrosion products has included some improvements which are based on the collaboration of the group of university of a coruña with the group of bertrand fritz at the cnrs and université de strasbourg (france). The model has been improved by accounting for: 1) fe sorption via cation exchange and surface complexation in three types of sorption sites; 2) kinetically-controlled corrosion rate. the second case deals with the radionuclide transport in the near field of a spent fuel repository in a granite formation. Current performance assessment models (pam) for radionuclide migration through the near field of a high-level radioactive waste repository usually rely on simplifying assumptions such as the use of the ‘kd approach’ for nuclide sorption and ‘the limited solubility’ for nuclide precipitation. Testing the validity of these assumptions has been limited by the lack of: 1) data on nuclide surface complexation and cation exchange; and 2) computer codes which could solve for the migration, sorption and precipitation of radionuclides simultaneously with the geochemical evolution of the near field of a hlw repository. Lab experiments performed in recent years have provided substantial data and understanding on the mechanisms of nuclide sorption. On the other hand, sophisticated research-oriented process-based computer codes and models have been developed which allow for the simultaneous modelling of migration, sorption, nuclide precipitation and multicomponent geochemical evolution of the near field. The validity of the ‘kd approach’ assumption for nuclide sorption and precipitation in the 0.75 m thick compacted bentonite barrier of a spent-fuel repository in granite has been evaluated and tested. Such testing has been performed by comparing the results of a typical pam with those obtained with a multicomponent reactive transport model which incorporates a mechanistic thermodynamic sorption model and has been solved with the updated version of core. Models have been performed for radionuclides for which detailed sorption data were available. They include ni2+ and u. Model results indicate that when the concentration of dissolved ni2+ is smaller than 10-6 mol/l, the kd of ni2+ remains constant and most of the ni2+ is sorbed on strong sorption sites. With the increase of the concentration of dissolved ni2+, strong sorption sites are saturated and the concentration of ni2+ sorbed on weak sorption sites increases and becomes the main component of sorbed ni2+. Exchanged ni2+ is 10 times smaller than ni2+ sorbed on weak sorption sites. The migration of the ni2+ front causes a slight decrease of ph. The kd of ni2+ is sensitive to the boundary water flux, the diffusion coefficient of the bentonite and the concentration of weak sorption sites. The kd of ni2+ is mainly controlled by ph and the concentration of dissolved ni2+. The competition effect of cs+ on ni2+ sorption is not significant because cs+ sorbs by cation exchange while ni2+ mainly sorbs by surface complexation. The corrosion products change the ph and eh in the bentonite significantly and affect significantly the sorption of ni2+ in the bentonite. The results of the tsm have been compared with those of classical models based on the use of a kd. A constant kd model fails greatly to reproduce the release rates of ni2+ from the bentonite into the granite. A model with a variable kd which depends on the concentration of dissolved ni2+ improves the results of the release rates. The best results of a kd model are obtained when the kd depends on the concentration of dissolved ni2+ and also on ph. the concentration of dissolved u in the bentonite barrier is very small under reducing conditions. The selection of the sorbing uranium species is very important for the simulation of uranium transport and sorption. The form of u, u(iv) or u(vi), in the spent fuel is also relevant for u migration. If u in the canister is assumed to be present as a u(iv) oxide, the concentration of dissolved u is very small. On the contrary, when a u(vi) mineral is considered in the canister, the concentration of dissolved u is too large. Most of the u sorbs on strong sites. The concentration of exchanged u is very small. Kd of u increases when the sorption of u4+ and uranyl carbonates species is considered. Most of the available u thermodynamic sorption data correspond to u(vi) and data for u(iv) are limited. The model of u sorption has revealed that there are uncertainties in: 1) the redox state and the form in which u is present in the spent fuel. Although uo2(am) is the most likely mineral phase, other mixed forms such as u4o9, u3o7 and u3o8 cannot be discarded; 2) the mineral phases controlling u solubility at the conditions of the repository. Such mineral phases have been analyzed with the help of eh-ph-u solubility plots, but there are uncertainties in the controlling phases; 3) the thermodynamic data for some u mineral phases such as coffinite; 4) the thermodynamic sorption data for u(vi) and u(iv) which are not comprehensive. Data for the sorption of uranyl carbonates are lacking. There is a need for a complete sorption data for multisite sorption materials. the last application of core2dv4 corresponds to the multicomponent reactive transport model of the water quality of the as pontes mine lake near a coruña in northwest spain. Similar to the previous model of this lake carried out by vattenfal (2003), the model performed in this dissertation assumes that lake waters are fully mixed throughout the four years during which the lake is being filled with: 1) rainfall on the surface of the lake and surface runoff on its drainage basin, 2) runoff of the nearby mine dumps and 3) fresh-water diverted from the eume river. Motivated by the needs of the owner of the mine, endesa, the model developed in this dissertation has been designed to: 1) replicate the predictions of the chemical evolution of vattenfal (2003); and 2) update the predictions of the evolution of lake chemistry during lake creation with core by accounting for the most recent chemical data. Prediction uncertainties of ph and concentrations of sulfate, mn, fe and al of lake waters have been evaluated by a sensitivity analysis with respect to: 1) hydrological data for water inflows; 2) chemical composition of mine surface runoff; 3) chemical composition and ph of surface runoff of mine dumps; 4) the solubility of ferrihydrite; and 5) the amount of runoff of mine water. The effect of several remediation measures on the final water quality has been evaluated. Model results show that the addition of cao(s) and caco3(s) to the inflow water can neutralize the lake water. Reducing the runoff of acid-producing mine areas by tapping them with a clay cover has also beneficial effects on the final water quality. The alkaline effluents coming from the nearby as pontes power plants play also a very important role to improve the conditions of the lake water.
Datos académicos de la tesis doctoral «Reactive transport models in nuclear waste disposal and acid mine drainage«
- Título de la tesis: Reactive transport models in nuclear waste disposal and acid mine drainage
- Autor: Chuanhe Lu
- Universidad: A coruña
- Fecha de lectura de la tesis: 11/12/2009
Dirección y tribunal
- Director de la tesis
- Javier Samper Calvete
- Tribunal
- Presidente del tribunal: bertrand Fritz
- Antonio Paz gonzalez (vocal)
- wilfried Pfingsten (vocal)
- johannes Lí¼tzenkirchen (vocal)